docker-expert
Pass
Audited by Gen Agent Trust Hub on Jul 31, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill utilizes standard Docker CLI tools (docker info, docker ps, docker build) and Unix utilities (find, grep) to analyze the container environment. These are routine operations for a DevOps-focused tool.
- [SAFE]: All external container images referenced (e.g., node:18-alpine, postgres:15-alpine, and Google's Distroless images) are sourced from official and well-known repositories.
- [COMMAND_EXECUTION]: The skill performs local environment discovery and validation using shell commands. This includes building and running temporary containers (docker build, docker run) to verify configurations, which is necessary for its primary function.
- [PROMPT_INJECTION]: The skill includes an indirect prompt injection attack surface because it ingests untrusted data from the local file system (via file discovery) and interpolates it into the agent context.
- Ingestion points: Local project structure and Dockerfile content analyzed via shell.
- Boundary markers: Absent in command output interpolation.
- Capability inventory: Shell execution via docker build, docker run, and docker exec.
- Sanitization: None implemented for external file content.
- [SAFE]: No evidence of data exfiltration or unauthorized network communication was detected. The instructions prioritize security hardening, such as dropping capabilities and managing secrets via Docker Secrets rather than environment variables.
Audit Metadata